Role of Asymmetrical Dimethylarginine in Renal Microvascular Endothelial Dysfunction in Chronic Renal Failure with Hypertension

181

Hypertens Res Vol.28 (2005) No.2 p.181 Original Article

Ken OKUBO, Koichi HAYASHI, Shu WAKINO, Hiroto MATSUDA, Eiji KUBOTA, Masanori HONDA, Hirobumi TOKUYAMA, Tokunori YAMAMOTO*, Fumihiko KAJIYA**, and Takao SARUTA

We examined whether endothelial function of the renal microcirculation was impaired in a model of chronic renal failure (CRF), and further assessed the role of asymmetrical dimethylarginine (ADMA) and its degrading enzyme, dimethylarginine dimethylaminohydrolase (DDAH), in mediating the deranged nitric oxide (NO) synthesis in CRF. CRF was established in male mongrel dogs by subtotal nephrectomy, and the animals were used in experiments after a period of 4 weeks. The endothelial function of the renal afferent and efferent arterioles was evaluated according to the response to acetylcholine, using an intravital needle-lens charge-coupled device camera. Intrarenal arterial infusion of acetylcholine (0.01 µg/kg/min) elicited 22±2% and 20±2% dilation of the afferent and efferent arterioles in normal dogs. In dogs with CRF, this vasodilation was attenuated (afferent, 12±2%; efferent, 11±1%), and the attenuation paralleled the diminished increments in urinary nitrite+nitrate excretion. In the animals with CRF, plasma concentrations of homocysteine (12.2±0.7 vs. 6.8±0.4 µmol/l) and ADMA were elevated (2.60±0.13 vs. 1.50±0.08 µmol/l). The inhibition of S-adenosylmethionine-dependent protein arginine N-methyltransferase by adenosine dialdehyde decreased plasma ADMA levels, and improved the acetylcholine-induced changes in urinary nitrite+nitrate excretion and arteriolar vasodilation. Acute methionine loading impaired the acetylcholine-induced renal arteriolar vasodilation in CRF, but not normal dogs, and the impairment in CRF dogs coincided with the changes in plasma ADMA levels. Real-time polymerase chain reaction revealed downregulation of the mRNA expression of DDAH-II in the dogs with CRF. Collectively, these results provide direct in vivo evidence of endothelial dysfunction in canine CRF kidneys. The endothelial dysfunction was attributed to the inhibition of the NO production by elevated ADMA, which involved the downregulation of DDAH-II. The deranged NO metabolic pathway including ADMA and DDAH is a novel mechanism for the aggravation of renal function. (Hypertens Res 2005; 28: 181–189)

Key Words: asymmetrical dimethylarginine, renal disease, nitric oxide, renal microcirculation, endothelium

mortality in patients with chronic renal failure (CRF) (1, 2), per se Introduction and renal disease increases the incidence of cardiovascular events even in the presence of macro- (3) or microalbu- Cardiovascular disease is a major cause of morbidity and minuria (4). Various factors have been identified as risk

From the Department of Internal Medicine, School of Medicine, Keio University, Tokyo, Japan; *Department of Urology, Nagoya University Graduate School of Medicine, Nagoya, Japan; and **Department of Cardiovascular Physiology, Okayama University Graduate School of Medicine and Dentistry, Okayama, Japan. Address for Reprints: Koichi Hayashi, M.D., Ph.D., Department of Internal Medicine, School of Medicine, Keio University, 35 Shinanomachi, Shin- juku-ku, Tokyo 160−8582, Japan. E-mail: [email protected] Received December 24, 2003; Accepted in revised form November 2, 2004. 182 Hypertens Res Vol. 28, No. 2 (2005)

Table 1. Baseline Hemodynamic and Blood Parameters in Table 2. Baseline Characteristics of Dogs with Normal Kid- Dogs with Normal Kidneys and Chronic Renal Failure neys and Chronic Renal Failure Control CRF Control CRF n 77 N 77 Body weight (kg) 12.2±1.1 10.4±1.0* von Willebrand factor (%) 11.3±0.8 31.0±8.0* MAP (mmHg) (conscious) 96±3 119±2** PAI-1 (nmol/l) 352±127 848±190# HR (beats/min) (conscious) 114±495±8* Homocysteine (μmol/l) 6.8±0.4 12.2±0.7** Hematocrit (%) 49±239±2** ADMA (μmol/l) 1.50±0.08 2.60±0.13** μ ± ± Serum creatinine ( mol/l) 64.6 2.7 188.5 26.5** CRF, chronic renal failure; PAI-1, plasminogen activator inhibi- ± ± Total cholesterol (mmol/l) 3.25 0.05 3.35 0.25 tor-1; ADMA, asymmetrical dimethylarginine. Data are HDL-cholesterol (mmol/l) 2.69±0.11 2.77±0.12 expressed as the mean±SEM. *p<0.05 vs. controls. **p<0.01 Triglyceride (mmol/l) 0.21±0.03 0.38±0.05* vs. controls. #p=0.05 vs. controls. Creatinine clearance (ml/min) 31.3±1.5 13.2±0.8** CRF, chronic renal failure; MAP, mean arterial pressure; HR, heart rate; HDL, high density lipoprotein. Data are expressed as these enzymes have not been elucidated in CRF. the mean±SEM. *p<0.05 vs. controls. **p<0.01 vs. controls. In the present study, we directly examined the endothelial function of the renal microvasculature in CRF by using an intravital needle-lens charge-coupled device (CCD) camera factors for cardiovascular events in CRF, including a modest technique, which allowed direct visualization of the renal increase in blood pressure (5, 6), left ventricular hypertrophy microvessels in an in vivo, in situ, and relatively intact exper- (7), and elevated lipoprotein (a) (8). Recently, hyperho- imental setting (20, 21). We observed that the acetylcholine mocysteinemia has been recognized as an independent risk (ACh)-induced vasodilation of the renal microcirculation and factor promoting atherosclerosis by inducing endothelial dys- the NO utilization were impaired in CRF. Furthermore, we function (9), and thus constitutes a potential cardiovascular observed an elevation in plasma ADMA and suppressed risk factor in CRF (10). Although the mechanism for hyper- expression of the renal DDAH-II mRNA level in CRF. The homocysteinemia-induced cardiovascular disease still pharmacological inhibition of PRMT by adenosine dialde- remains controversial, several lines of studies suggest that hyde restored the elevated plasma ADMA level and the homocysteine limits the bioavailability of NO (11) and blunted microvascular response to ACh. These findings sug- increases oxidative stress (12), thereby favoring hyper-reac- gest that the accumulation of ADMA by the suppressed tivity of vascular beds (13). It is surmised therefore that in DDAH-II expression within the kidney is one possible mech- CRF, endothelial dysfunction within the renal microvascula- anism for endothelial dysfunction in CRF. ture accelerates the renal injury, which would constitute a vicious circle for the progression of end-stage CRF. Never- Methods theless, the role of homocysteine and the disturbed NO production in mediating the development of renal microvascular All experimental procedures were conducted according to the dysfunction has not been elucidated. institutional guidelines of the Animal Care Committee. Fifty A growing body of evidence has accrued that CRF is asso- adult male mongrel dogs (9−14 kg) were fed a standard diet ciated with deranged NO metabolism. Valance et al. (14) (Oriental Yeast Co., Tokyo, Japan) and were divided into an originally demonstrated the presence of an endogenous NO ACh protocol group (normal controls: n=7; CRF: n=7), an synthase inhibitor, asymmetric dimethylarginine (ADMA), ACh+adenosine dialdehyde protocol group (normal controls: and subsequent studies have shown that accumulation of n=7; CRF: n=7), a methionine loading protocol group (nor- ADMA is an important risk factor for cardiovascular disease mal controls: n=6; CRF: n=6), and a protocol group for the in CRF (15, 16). ADMA is biosynthesized from arginine by quantification of DDAH mRNA expression (normal controls: S-adenosylmethionine-dependent protein arginine N-methyl- n=5; CRF: n=5). Dogs in the CRF groups were anesthetized transferase (PRMT), which transmethylates S-adenosylme- with sodium pentobarbital (25 mg/kg), and CRF was induced thionine, an intermediate in the conversion of methionine to by ligation of two or three branches of the left renal artery and homocysteine (17). Furthermore, ADMA degradation is reg- contralateral nephrectomy. Dogs in the normal control groups ulated by dimethylarginine dimethylaminohydrolase were anesthetized with sodium pentobarbital (25 mg/kg) and (DDAH), which has recently been shown to have two iso- sham-operated. Four weeks after the operation, dogs of all forms, DDAH-I and DDAH-II, and exhibits different tissue groups underwent the same protocol. Animal preparations distributions and regulations for each isoform (18, 19). were detailed in our previous publications (20−22). An elec- Although both homocysteine and ADMA are suggested to be tromagnetic flow probe was placed around the left renal novel candidates causing endothelial dysfunction in CRF by artery for measurement of renal blood flow (RBF). impairing NO activity, the precise regulatory mechanisms for Renal microcirculation was assessed with an intravital nee- Okubo et al: Renal Endothelial Impairment in Renal Disease 183

Fig. 1. Visualization of renal microvessels and effects of acetylcholine on renal microvessels. The diameter of afferent arterioles (AFF) in chronic renal failure (CRF) dogs was greater than that in normal dogs, whereas no difference in efferent arteriolar (EFF) diameter was noted (A, B). Furthermore, the acetylcholine-induced vasodilation of AFF and EFF was blunted in CRF dogs (triangles, n=7) compared with that in normal dogs (circles, n=7; C, D). The arteriolar responses to acetylcholine were expressed as a percentage of nitroprusside-induced vasodilation. Results are expressed as the mean±SEM. SNP, sodium nitroprusside. **p<0.01 vs. baseline; †p<0.05, ††p<0.01 vs. normal dogs. ¶p<0.05 vs. 0.01 μg/kg/min acetylcholine. dle-lens CCD camera (VMS-1210; Nihon Kohden, Tokyo, tion of basal hemodynamics, ACh (0.01 μg/kg/min) was Japan) (20, 21). After the surgical procedure and instrumenta- administered into the renal artery. Following the 30-min tion, a CCD probe was introduced into the left kidney. The observation of the ACh-induced response, adenosine dialde- animals were allowed to equilibrate for 60 min before the ini- hyde (15 μg/kg/min; Sigma) was infused into the renal artery tiation of experimental protocols. and the effects of these agents were assessed. For the final protocol, dogs of both the CRF and normal kidney groups received L-methionine (100 mg/kg; Ajino- Protocols moto, Tokyo, Japan) orally at 7:00 in the morning. A standard After allowing the basal hemodynamics to stabilize for 20 breakfast was served 2 h after methionine administration. min, ACh (Sigma, St. Louis, USA) or sodium nitroprusside Plasma homocysteine and ADMA concentrations were (SNP; Maruishi Seiyaku, Tokyo, Japan) was infused directly assayed at 0, 5, 8, 12, and 24 h after methionine administra- into the renal artery at increasing doses of 0.0001 to 0.01 μg/ tion. Likewise, urinary NOx excretion and renal microvascu- kg/min for ACh, and 0.002 μg/kg/min for SNP. The effects of lar responses to ACh (0.01 μg/kg/min) were evaluated at 0, 5, these agents on RBF, urinary nitrite+nitrate (NOx) and the and 12 h after methionine administration. renal microvascular tone were evaluated. A 30-min interval after each ACh dose was required for stabilization of the Biochemical Analyses response. To examine whether the ADMA-mediated NO inhibition Plasma homocysteine and ADMA were determined by high was responsible for the impaired renal endothelial response, performance liquid chromatography (23, 24). Urinary NOx the effects of adenosine dialdehyde, a PRMT inhibitor, on the concentrations were evaluated using the Griess reaction (21, plasma ADMA concentration and the ACh induced-response 25). of renal microcirculation were evaluated. After the stabiliza- 184 Hypertens Res Vol. 28, No. 2 (2005)

Fig. 2. Effects of acetylcholine on renal blood flow (RBF), and urinary nitrate/nitrite excretion. Acetylcholine-induced increases in RBF and urinary nitrate+nitrite (NOx) excretion were smaller in dogs with chronic renal failure (CRF; triangles, n=7) than in normal dogs (circles, n=7). Results are expressed as the mean±SEM. SNP, sodium nitroprusside. *p<0.05, **p<0.01 vs. baseline. †p<0.05, ††p<0.01 vs. normal dogs. ¶p<0.05 vs. 0.01 μg/kg/min acetylcholine.

Fig. 3. Effects of adenosine dialdehyde on plasma asymmetrical dimethylarginine, homocysteine, urinary nitrate/nitrite excretion, and renal hemodynamics. Plasma asymmetrical dimethylarginine (ADMA) and homocysteine were elevated in dogs with chronic renal failure (CRF). Adenosine dialdehyde (AD; 15 μg/kg/min) suppressed plasma ADMA, but not homocysteine. In dogs with normal kidneys, AD had no effect on the changes in urinary nitrate+nitrite (NOx) excretion, renal blood flow (RBF), or renal arteriolar diameter induced by intrarenal infusion of acetylcholine (ACh; 0.01 µg/kg/min). In CRF dogs, the ACh- induced changes in these parameters were diminished, and were partially restored by the pretreatment with AD. Results are expressed as the mean±SEM (n=7). *p<0.05 vs. baseline. †p<0.05, ††p<0.01 vs. normal dogs.

the quantification of mRNA expression of DDAH-I/II, real- mRNA Isolation, cDNA Synthesis, and Real-Time time PCR was performed using the ABI PRISM-7700 Polymerase Chain Reaction (PCR) sequence detector (PE Applied Biosystems, Tokyo, Japan). Total RNA was isolated from the canine renal cortex with SYBER Green I Dye (PE Applied Biosystems) was utilized Trizol Reagent (Invitrogen, Carlsbad, USA). Total RNA (50 for detection of the PCR reaction. Each set of primers yielded ng) was reverse-transcribed for cDNA synthesis with the a single amplified PCR product with a sequence identical to SuperScript First-Strand Synthesis System (Invitrogen). For one of those published in GenBank. Okubo et al: Renal Endothelial Impairment in Renal Disease 185

Fig. 4. Temporal profiles of the effects of oral methionine loading on plasma concentrations of homocysteine and asymmetrical dimethylarginine, urinary nitrate/nitrite excretion and blood pressure. Acute methionine loading evoked prominent increases in plasma homocysteine levels in dogs with normal and chronic renal failure (CRF) kidneys, whereas a marked elevation in plasma asymmetrical dimethylarginine (ADMA) level was seen in CRF dogs. Urinary nitrate+ nitrite (NOx) excretion was decreased only in CRF dogs. In parallel, exaggerated elevations in blood pressure were observed in CRF dogs. Results are expressed as the mean±SEM (n=6). *p<0.05, **p<0.01, ***p<0.001 vs. 0 h. †p<0.05, ††p<0.01 vs. normal dogs.

Statistics Effect of ACh and SNP on Renal Microcirculation Results are expressed as the mean±SEM. Data were analyzed Using an intravital CCD camera, we examined whether CRF by 1-way/2-way ANOVA, as appropriate, followed by Bon- altered the endothelial function of renal afferent and efferent ferroni’s post hoc test (StatView, SAS Institute Inc., Cary, arterioles in vivo. In normal dogs, the basal diameters of the USA). Values of p<0.05 were considered statistically signif- afferent and efferent arterioles were 14.4±0.2 (n=7) and icant. 12.5±0.5 μm (n=7), respectively (Fig. 1A). Direct intrarenal infusion of ACh caused dose-dependent dilation of the affer- ± ± Results ent and efferent arterioles, with 22 2% and 20 2% increments in diameter observed at 0.01 μg/kg/min, respectively. Further administration of SNP had no effect on either afferent Baseline Characteristics of CRF Dogs or efferent arterioles. Table 1 shows baseline characteristics of the dogs with CRF In CRF dogs, the basal afferent arteriolar diameter was and the normal controls. After 4 weeks of five-sixths nephre- greater (24.7±2.2 μm, n=7) than that in normal controls ctomy, body weight, heart rate, and glomerular filtration rate (p<0.01; Fig. 1B), but no difference in efferent arteriolar were decreased. Mean arterial pressure did not change at 2 diameter was noted between CRF (13.0±0.8 μm, n=7) and weeks of subtotal nephrectomy (104±6 mmHg, p>0.1, n=7), normal dogs (p>0.5). Furthermore, the vasodilator response but a significant elevation was observed at 4 weeks. Serum to ACh was diminished in CRF dogs; at 0.01 μg/kg/min, only creatinine levels were elevated from 64.6±2.7 to 188.5±26.5 12±2% and 11±1% increments were observed in the afferent μmol/l. Although triglyceride levels were elevated, total cho- and efferent arterioles, respectively. The subsequent addition lesterol and high density lipoprotein (HDL)-cholesterol levels of SNP dilated both arterioles by an increment of 16±2% and did not change. Plasma levels of von Willebrand factor and 18±2% of the baseline value, respectively. When the arteri- plasminogen activator inhibitor-1 (PAI-1), markers for endo- olar responses to ACh were expressed as a percentage of the thelial function, were elevated in the dogs with CRF (Table SNP-induced vasodilation, it was evident that the responses 2). Both plasma homocysteine and ADMA concentrations in CRF were blunted compared with those in normal dogs were increased. Plasma ADMA concentrations were elevated (Fig. 1C and D). The vasodilator response to ACh (0.01 μg/ at 2 weeks of five-sixths nephrectomy (2.20±0.14 μmol/l, kg/min) was also diminished at 2 weeks of subtotal nephrec- p<0.01, n=7). tomy (afferent, 15±1% increments, p=0.001 vs. normal controls; efferent, 14±4% increments, p=0.03 vs. normal 186 Hypertens Res Vol. 28, No. 2 (2005)

Fig. 5. Sequential changes in acetylcholine-induced actions on renal arterioles and urinary nitrate/nitrite excretion during oral methionine loading. Oral methionine loading had no effect on the arteriolar responses to acetylcholine (ACh) or urinary nitrate+nitrite (NOx) excretion in normal dogs. In contrast, in dogs with chronic renal failure (CRF), the ACh-induced dilation of the renal arterioles and the increase in urinary NOx excretion were markedly attenuated at 5 h of acute methionine loading, and these attenuations tended to persist at 12 h. Results are expressed as the mean±SEM (n=6).

controls). diameters of afferent and efferent arterioles were enhanced by We further investigated the effects of ACh on RBF and uri- adenosine dialdehyde. nary NOx excretion in CRF (Fig. 2). Baseline levels of RBF (61±6 ml/min, n=7) and urinary NOx excretion (11.0±2.4 Effects of Oral Methionine Loading on Plasma nmol/min, n=7) in the animals with CRF were less than those Homocysteine and ADMA Concentrations in normal controls (RBF, 185±11 ml/min, n=7, p<0.01; urinary NOx, 32.2±11.0 nmol/min, n=7, p<0.01). In normal Acute oral methionine loading markedly elevated the plasma dogs, RBF and urinary NOx excretion were augmented in a homocysteine levels in both normal and CRF dogs (Fig. 4). In dose-dependent manner; at 0.01 μg/kg/min, ACh elicited contrast, the methionine administration caused only a modest 28±3% increments in RBF and 239±56% increments in uri- elevation in the plasma ADMA concentration in normal dogs, nary NOx excretion. In CRF dogs, blunted ACh-induced whereas a prominent increase in the plasma ADMA concen- increases in RBF and urinary NOx excretion were observed. tration was noted in CRF dogs. Similarly, methionine loading had only a modest, nonsignificant effect on urinary NOx excretion in normal dogs, but a significant decrease was Effects of PRMT Inhibition on Renal Hemodynam- observed 5 h after methionine loading in CRF dogs. Concom- ics and Microcirculation itantly, blood pressure was elevated in both normal and CRF In normal dogs, the administration of adenosine dialdehyde dogs, although the changes were greater in CRF dogs. had no effect on plasma ADMA/homocysteine levels (Fig. 3), In parallel with the changes in plasma ADMA concentra- urinary NOx excretion, RBF or renal arteriolar diameters tions, the ACh (0.01 μg/kg/min)-induced vasodilation of (n=7). Furthermore, adenosine dialdehyde did not alter the renal afferent and efferent arterioles was markedly impaired ACh-induced changes in these parameters (n=7, Fig. 3). In in CRF dogs at 5 h of oral methionine loading, and this ten- CRF dogs, however, adenosine dialdehyde caused a signifi- dency persisted at 12 h (p=0.1 vs. 0 h; Fig. 5). In contrast, in cant decrease in plasma ADMA concentration (from normal dogs, methionine had no effect on the ACh-induced 2.58±0.11 to 2.23±0.10 μmol/l, n=7, p<0.05), and increases vasodilation of these arterioles. These results paralleled the in urinary NOx excretion (from 25.0±6.2 to 61.0±7.1 nmol/ changes in the urinary NOx excretion induced by ACh; the min, n=7, p<0.05), RBF (from 66±2 to 76±2 ml/min, ACh-induced increments in urinary NOx excretion were p<0.05) and renal arteriolar diameters (afferent, from markedly attenuated at 5 h of methionine loading in CRF 26.4±0.6 to 28.6±0.5 μm, p<0.05; efferent, from 15.0±0.2 dogs, whereas only a modest inhibitory tendency was seen in to 16.5±0.3 μm, p<0.05). Plasma homocysteine levels normal dogs. tended to decrease, although the changes did not attain statistical significance. Expression of DDAH-I/II in Normal and Remnant In CRF dogs, the diminished ACh (0.01 μg/kg/min)- Kidneys induced increases in urinary NOx excretion (from 11.0±2.4 to 25.0±6.2 nmol/min, p<0.05, n=7) was prominently The mRNA expression level of each isoform of DDAH in the restored by adenosine dialdehyde (to 61.0±7.1 nmol/min) kidney was assessed by real-time PCR. The mRNA levels of (Fig. 3). Similarly, the ACh-induced changes in RBF and the DDAH-I tended to be downregulated in CRF, although the Okubo et al: Renal Endothelial Impairment in Renal Disease 187

dysfunction in CRF has been documented in various experimental settings (26, 27), and indeed is anticipated from the increased plasma von Willebrand factor and PAI-1 (Table 2), no investigations have been conducted examining the endothelial function at the renal microvascular level in vivo. The present study thus offers the first direct evidence for the impaired endothelial function of the renal microvasculature in CRF. It is worth noting that the afferent arterioles in CRF dogs have larger diameters than those in normal dogs. Although the mechanism of the afferent arteriolar dilation in CRF remains unestablished, the diminished tone of this vessel would pro- Fig. 6. mRNA expression of dimethylarginine dimethylami- mote glomerular hypertension and the subsequent develop- nohydrolase (DDAH) in the kidney. The mRNA expression of ment of renal injury. Furthermore, an increase in the DDAH-I and DDAH-II in the normal kidneys and the rem- diameters of the afferent arterioles of CRF dogs may augment nant kidneys with chronic renal failure (CRF) was analyzed the wall tension of these arterioles, and thereby facilitate by real-time polymerase chain reaction. The ratio between endothelial dysfunction. Nevertheless, the diminished the mRNA levels of each isoform of DDAH and the mRNA response of the efferent arterioles to ACh clearly indicates a level of GAPDH, a constitutively expressed isoform, was CRF-induced alteration in the responsiveness of this vessel used for the comparison. The sequences for the forward and segment. reverse primers were: 5′-catggctgggcctaacctaat-3′ and 5′- Among various factors responsible for the endothelial dys- tgagtttgtcatagcggtggtc-3′ for DDAH-I, 5′-aggtaccagggt function in CRF, we have focused on two substances that are ′ ′ ′ gacatcagaga-3 and 5 -cagctgctgactgcctctttc-3 for DDAH-II, reported to contribute to the regulation of vascular tone in ′ ′ ′ and 5 -ggacaaatcaacgaggtgct-3 and 5 -ggacaaatcaacgaggt various vascular beds, ADMA and homocysteine. ADMA has gct-3′ for glyceraldehyde-3-phosphate dehydrogenase been shown to inhibit NO synthase by competing with L-argi- (GAPDH), respectively. The amplification conditions were nine (28), and plasma ADMA levels are elevated in both 95°C for 10 min, followed by 40 cycles of 95°C for 15 s and patients (14, 16) and experimental animals with CRF (Table 60°C for 60 s. Results are expressed as the mean±SEM 2). It has been demonstrated that elevated plasma ADMA (n=5). n.s., not significant. concentrations are significantly related to the degree of impaired endothelium-dependent forearm vasodilation (29). Indeed, the present study showed that the reduction in the difference did not attain statistical significance (Fig. 6). In plasma ADMA level by adenosine dialdehyde (17) enhanced contrast, the mRNA levels of DDAH-II were markedly down- the urinary NOx excretion and restored the ACh-induced regulated in the dogs with CRF (p<0.01 vs. normal controls, changes in renal hemodynamic parameters, including RBF n=5). and afferent and efferent arteriolar diameters in dogs with CRF, but not in those with normal renal function (Fig. 3). Discussion Conversely, acute ADMA elevation by methionine loading was associated with elevated blood pressure (Fig. 4) and a In the present study, we have demonstrated that the endothe- blunted vasodilator response to ACh in CRF dogs (Fig. 5). lium-dependent vasodilation of renal microvessels is ADMA therefore constitutes an important factor contributing impaired in subtotally nephrectomized CRF dogs, with the to the derangement in endothelial function in renal microcir- use of a novel technique that directly visualizes the in vivo, in culation, and may also be responsible for the acceleration of situ, and relatively intact renal microcirculation, i.e., an intra- cardiovascular events in CRF (16). In this regard, we also vital needle-lens CCD camera technique (20, 21). The alter- found elevated plasma ADMA concentrations (2.20±0.14 ation in the endothelium-dependent vasodilation of renal μmol/l) and the diminished vasodilator response to ACh at 2 arterioles is closely linked to the plasma ADMA and NO pro- weeks of five-sixths nephrectomy, at which stage mean arte- duction, rather than the homocysteine concentration. Finally, rial pressure was unaltered (104±6 mmHg). In concert, it is the modification of the plasma ADMA level by the blockade conjectured that elevated plasma ADMA concentrations and of PRMT activity or methionine loading alters the endothelial endothelial dysfunction are attributable to the changes in function in CRF, which parallels the changes in urinary NOx internal milieu induced by CRF, although the elevated blood excretion. Collectively, these observations strongly suggest pressure per se may also modify ADMA production at an important contribution of ADMA, an endogenous NO syn- advanced stages of CRF. It is of note that the plasma triglyc- thesis inhibitor, to the renal microvascular abnormality in eride level is greater in the dogs with CRF than in the normal CRF, which would account in part for the deranged renal and/ controls in the present study, and acute hypertriglyceridemia or systemic NO production in CRF. Although endothelial following high fat diet intake has been reported to be associ- 188 Hypertens Res Vol. 28, No. 2 (2005) ated with endothelial impairment (30). Nevertheless, it has of mRNA is consistent with the change in the enzyme activ- also been demonstrated that chronic hypertriglyceridemia ity; acute methionine loading elicited a marked elevation in does not affect endothelial function (31). Further studies are ADMA concentration in CRF, whereas in normal controls required to clarify this issue. only a modest increase in the plasma ADMA level was noted The present study also demonstrated an increased level of (Fig. 4). Taken together, our studies indicate that the down- plasma homocysteine in CRF dogs. Previous studies have regulation of DDAH-II constitutes a critical determinant of shown that hyperhomocysteinemia is associated with the CRF-induced impairment of renal microcirculation. impaired endothelial function in animals and humans (12, In this regard, Stühlinger et al. (35) have recently demon- 28). Although the precise mechanisms of the impairment are strated that homocysteine impairs the NO synthase pathway not fully determined, a direct NO-quenching effect by oxida- through the inhibition of DDAH in cultured endothelial cells. tive stress is one possible mechanism (12). Recently, Böger et Whereas their findings observed in vitro show an inhibitory al. (32) reported that acute hyperhomocysteinemia induced action of homocysteine on DDAH activity and subsequently a by methionine administration increases plasma ADMA con- positive correlation between homocysteine and ADMA lev- centrations, and impairs the flow-mediated vasodilation of the els, our observations obtained in normal dogs indicate disso- brachial artery in healthy human subjects. They suggested ciation between plasma homocysteine and ADMA levels. that the impaired vasodilation during the acute homocysteine These discrepant findings may be attributed to the different elevation may be attributed to an increase in ADMA. In the experimental settings (e.g., in vivo vs. in vitro). Furthermore, present study, we examined the effect of acute methionine it is plausible that additional mechanisms may contribute to loading on plasma homocysteine and ADMA levels and the suppressed DDAH activity and expression in CRF. simultaneously assessed the ACh-induced vasodilation of Intriguingly, our results indicate that each DDAH isoform is renal arterioles. Our results showed that, although nearly the differently regulated in the kidney, a finding consistent with same level of plasma homocysteine was achieved in normal the previous report in pulmonary hypertension (19). Further and CRF dogs during the methionine administration, the studies will be needed to elucidate the regulatory mechanism impairment in the ACh-induced renal arteriolar vasodilation of DDAH-II in CRF. by methionine was associated with the level of ADMA, but In conclusion, the present study provides direct evidence of not homocysteine (Figs. 4 and 5). Furthermore, the adminis- the endothelial dysfunction in kidneys of CRF dogs in vivo, tration of adenosine dialdehyde failed to alter plasma with the use of an intravital needle-lens CCD camera tech- homocysteine levels, but decreased plasma ADMA and ame- nique. The impaired endothelial function in CRF was attribut- liorated the renal hemodynamic responses to ACh (Fig. 3). able to the NO synthase inhibition by ADMA, in which the Taken together, these observations favor the premise that downregulation of an ADMA-degrading enzyme, DDAH, ADMA plays a substantial role in mediating the impaired was involved, and possibly to elevated systemic blood pres- endothelial function in CRF. sure. Such deranged NO metabolism would aggravate renal Although the present study clearly indicates that ADMA function as well as increase the risk for cardiovascular events plays an important role in the endothelial dysfunction in CRF, in CRF. the regulatory mechanism of the plasma ADMA level has not been fully explored. It is reasonable to speculate that the ele- References vated ADMA concentration is attributable to inadequate renal clearance (33). Alternatively, as demonstrated in the present 1. Foley RN, Parfrey PS: Cardiovascular disease and mortality study, the metabolism of ADMA may be altered in CRF. in ESRD. J Nephrol 1998; 11: 239−245. Thus, the mRNA of DDAH-II, an ADMA-degrading enzyme, 2. Lindner A, Charra B, Sherrard DJ, Scribner BH: Accelerated is downregulated in kidneys from CRF dogs (Fig. 6), whereas atherosclerosis in prolonged maintenance hemodialysis. N − in our preliminary experiments no change was observed in the Engl J Med 1974; 290: 697 701. 3. Samuelsson O, Wilhelmsen L, Elmfeldt D, et al: Predictors level of PRMT, an ADMA-producing enzyme. Immunohis- of cardiovascular morbidity in treated hypertension: results tochemical analysis has shown that DDAH is widely distrib- from the primary preventive trial in Goteborg, Sweden. J uted within the rat kidney, including the endothelium in the Hypertens 1985; 3: 167−176. glomeruli and renal vasculature as well as the proximal 4. Diercks GF, Stroes ES, van Boven AJ, et al: Urinary albu- tubules and macula densa (34). Recently, Leiper et al. (18) min excretion is related to cardiovascular risk indicators, not cloned two isoforms of human DDAH, and found that to flow-mediated vasodilation, in apparently healthy sub- DDAH-I predominates in tissues expressing neuronal NOS, jects. Atherosclerosis 2002; 163: 121−126. whereas DDAH-II is distributed in the regions where endo- 5. Stefanski A, Schmidt KG, Waldherr R, Ritz E: Early thelial NOS is co-localized. Therefore, decreased expression increase in blood pressure and diastolic left ventricular mal- of DDAH-II in the renal vasculature of the remnant kidney function in patients with glomerulonephritis. Kidney Int − might result in the accumulation of ADMA, and this would 1996; 50: 1321 1326. 6. Hatta T, Nakata T, Harada S, et al: Lowering of blood pres- augment the inhibitory actions of ADMA on endothelial NOS sure improves endothelial dysfunction by increase of nitric in renal microvessels. Furthermore, an alteration at the level oxide production in hypertensive rats. Hypertens Res 2002; Okubo et al: Renal Endothelial Impairment in Renal Disease 189

25: 455−460. lens-probe videomicroscope. Kidney Int 2001; 60: 364−369. 7. Stack AG, Saran R: Clinical correlates and mortality impact 21. Matsuda H, Hayashi K, Arakawa K, et al: Zonal heterogene- of left ventricular hypertrophy among new ESRD patients in ity in action of angiotensin-converting enzyme inhibitor on the United States. Am J Kidney Dis 2002; 40: 1202−1210. renal microcirculation: role of intrarenal bradykinin. J Am 8. Kronenberg F, Kuen E, Ritz E, et al: Lipoprotein (a) serum Soc Nephrol 1999; 10: 2272−2282. concentrations and apolipoprotein (a) phenotypes in mild 22. Matsuda H, Hayashi K, Arakawa K, et al: Distinct modula- and moderate renal failure. J Am Soc Nephrol 2000; 11: tion of superficial and juxtamedullary arterioles by prosta- 105−115. glandin in vivo. Hypertens Res 2002; 25: 901−910. 9. Chauveau P, Chadefaux B, Coude M, et al: Hyperhomocys- 23. Vester B, Rasmussen K: High performance liquid chroma- teinemia, a risk factor for atherosclerosis in chronic uremic tography method for rapid and accurate determination of patients. Kidney Int 1993; 43 (Suppl 41): S72−S77. homocysteine in plasma and serum. Eur J Clin Chem Clin 10. Moustapha A, Naso A, Nahlawi M, et al: Prospective study Biochem 1991; 29: 549−554. of hyperhomocysteinemia as an adverse cardiovascular risk 24. Matsuoka H, Itoh S, Kimoto M, et al: Asymmetrical dimeth- factor in end-stage renal disease. Circulation 1998; 97: 138− ylarginine, an endogenous nitric oxide synthase inhibitor, in 141. experimental hypertension. Hypertension 1997; 29: 242− 11. Upchurch GR Jr, Welch GN, Fabian AJ, et al: Homocysteine 247. decreases bioavailable nitric oxide by a mechanism involv- 25. Ohta K, Araki N, Shibata M, et al: A novel in vivo assay sys- ing glutathione peroxidase. J Biol Chem 1997; 272: 17012− tem for consecutive measurement of brain nitric oxide pro- 17017. duction combined with the microdialysis technique. 12. Kanani PM, Sinkey CA, Browning RL, Allaman M, Knapp Neurosci Lett 1994; 176: 165−168. HR, Haynes WG: Role of oxidant stress in endothelial dys- 26. Wever R, Boer P, Hijmering M, et al: Nitric oxide produc- function produced by experimental hyperhomocyst(e)inemia tion is reduced in patients with chronic renal failure. Arterio- in humans. Circulation 1999; 100: 1161−1168. scler Thromb Vasc Biol 1999; 19: 1168−1172. 13. Okatani Y, Wakatsuki A, Reiter RJ: Protective effect of 27. Annuk M, Zilmer M, Lind L, Linde T, Fellstrom B: Oxida- melatonin against homocysteine-induced vasoconstriction of tive stress and endothelial function in chronic renal failure. J human umbilical artery. Biochem Biophys Res Commun Am Soc Nephrol 2001; 12: 2747−2752. 2000; 277: 470−475. 28. Tsikas D, Böger RH, Sandmann J, Bode-Böger SM, Frolich 14. Vallance P, Leone A, Calver A, Collier J, Moncada S: Accu- JC: Endogenous nitric oxide synthase inhibitors are respon- mulation of an endogenous inhibitor of nitric oxide synthesis sible for the L-arginine paradox. FEBS Lett 2000; 478: 1−3. in chronic renal failure. Lancet 1992; 339: 572−575. 29. Böger RH, Bode-Böger SM, Szuba A, et al: Asymmetric 15. Miyazaki H, Matsuoka H, Cooke JP, et al: Endogenous dimethylarginine: a novel risk factor for endothelial dysfunc- nitric oxide synthase inhibitor⎯a novel marker of athero- tion: its role in hypercholesterolemia. Circulation 1998; 98: sclerosis. Circulation 1999; 99: 1141−1146. 1842−1847. 16. Zoccali C, Bode-Böger S, Mallamaci F, et al: Plasma con- 30. Lundman P, Tornvall P: Acute triglyceridemia and endothe- centration of asymmetrical dimethylarginine and mortality in lial function. Am J Cardiol 2002; 90: 447. patients with end-stage renal disease: a prospective study. 31. Jonkers IJ, van de Ree MA, Smelt AH, et al: Insulin resis- Lancet 2001; 358: 2113−2117. tance but not hypertriglyceridemia per se is associated with 17. Böger RH, Sydow K, Borlak J, et al: LDL cholesterol upreg- endothelial dysfunction in chronic hypertriglyceridemia. ulates synthesis of asymmetrical dimethylarginine in human Cardiovasc Res 2002; 53: 496−501. endothelial cells: involvement of S-adenosylmethionine- 32. Böger RH, Lentz SR, Bode-Böger SM, Knapp HR, Haynes dependent methyltransferases. Circ Res 2000; 87: 99−105. WG: Elevation of asymmetrical dimethylarginine may medi- 18. Leiper JM, Santa Maria J, Chubb A, et al: Identification of ate endothelial dysfunction during experimental hyperho- two human dimethylarginine dimethylaminohydrolases with mocysteinaemia in humans. Clin Sci 2001; 100: 161−167. distinct tissue distributions and homology with microbial 33. Xiao S, Wagner L, Schmidt RJ: Circulating endothelial nitric arginine deiminases. Biochem J 1999; 343: 209−214. oxide synthase inhibitory factor in some patients with 19. Arrigoni FI, Vallance P, Haworth SG, Leiper JM: Metabo- chronic renal disease. Kidney Int 2001; 59: 1466−1472. lism of asymmetric dimethylarginines is regulated in the 34. Tojo A, Kimoto M, Wilcox CS: Renal expression of consti- lung developmentally and with pulmonary hypertension tutive NOS and DDAH: separate effects of salt intake and induced by hypobaric hypoxia. Circulation 2003; 107: angiotensin. Kidney Int 2000; 58: 2075−2083. 1195−1201. 35. Stühlinger MC, Tsao PS, Her JH, Kimoto M, Balint RF, 20. Yamamoto T, Hayashi K, Matsuda H, et al: In vivo visual- Cooke JP: Homocysteine impairs the nitric oxide synthase ization of tubuloglomerular feedback- and angiotensin II- pathway: role of asymmetric dimethylarginine. Circulation induced renal vasoconstriction in dogs by an intravital pencil 2001; 104: 2569−2575.